Sediment mixing rapidly and basin-wide cosmogenic nuclide analysis in eroding mountainous environments

Cosmogenic nuclide concentrations in alluvial sediments have been widely used to estimate basin-wide denudation rates. This technique requires that sediments be well mixed so as to avoid biasing results towards particular source locations within the basin. However, few studies have tested for well-mixed sediment with cosmogenic nuclide data. We examine sediment mixing with measurements of in situ-produced cosmogenic r Be in alluvium from small, high-relief catchments in the San Bernardino Mountains, California, USA. The mixing at the reach scale was tested with three samples of channel sediment taken at 5 m intervals. Adequate mixing is indicated with almost identical Be-10 concentrations for two samples while the third falls just outside one standard error. The mixing at channel confluences was tested at three sites where samples were obtained downstream of the tributary junction and from each of the tributaries upstream of the join. If the sediments are well-mixed, radionuclide concentrations for the downstream samples will reflect concentrations of the tributaries weighted by the rates of sediment production. The results show that sediment is insufficiently mixed at one junction, and indeterminate at another (that is, cosmogenic nuclide concentrations in the tributaries are too similar to determine source). At a third site, results suggest sufficient sediment mixing may be possible some distance downstream of the junction. Collectively, these findings demonstrate that the assumption of well-mixed sediments may be invalid in small drainages subject to episodic sediment delivery. The alluvial samples should be collected well downstream of tributary junctions and tests for sediment mixing should be performed. (C) 2006 Elsevier Ltd. All rights reserved.